Periodically, we update our version of the so-called “Fed Model” to provide a picture of how current US stock market valuation compares to Treasury yields from a historical perspective. Traditionally, the Fed Model takes the trailing twelve month P/E ratio for the S&P 500, inverts it to get what’s known as the “earnings yield” then takes the ratio of the earnings yield to the yield for the 10-year US Treasury bond. Theoretically, the higher the ratio, the more undervalued the market is and vice versa. There have been many justifiable criticisms of the traditional Fed Model, many due to the fact that the traditional Fed Model has been an imperfect predictor of future stock market out and under performance. We think some of this has to do with the traditional volatility of the trailing twelve month P/E ratio, which can jiggle violently during times of major movements in price and underlying earnings, especially when earnings completely fall out of bed.
To combat this issue, we substitute Shiller’s 10-year CAPE P/E for the trailing twelve month ratio as the “E” in the longer-term version is much more stable. The longer term chart for this indicator follows:
As you can see, using +/- 1 standard deviation bands as a guide, the model has performed reasonably well in signaling severe under and over valuation turning points, especially on the undervaluation side. A move above the range in early 1962 (undervaluation) presaged a strong move higher in markets for the next several years. Similarly, the move above the standard deviation range during the market lows of 1974 proved to be a decent point for what would become a bull market rally lasting until 2000, which is where extreme overvaluation levels were reached (though in fairness it took until 1982 for the bull market to begin in earnest). The model again signaled extreme undervaluation levels in the wake of the financial crisis in late 2008. Since that time, the model has remained in severe undervaluation territory, primarily due to the fact that 10-year Treasury yields remain so low. During that time, US markets have tripled off the lows on a nominal return basis and performed even better when dividends are taken into account.
As of now, owing to the continued low bond yields, the model remains in undervalued territory, despite the fact that long-term CAPE P/E ratio is reaching very extended levels. Many observers would argue that the model is “broken” due to the outsized influence of extraordinary Federal Reserve policy on long-term interest rates. Over the past several years, however, markets have continued to move higher validating the signal, even though it’s been “distorted.”
The signal could move out of undervalued territory one of several ways. A strong spike in Treasury yields would of course push the ratio down all things being equal. Or, for better or worse, the market could take off to the upside recreating the extreme CAPE valuation bubble scenario observed in 1999 and 2000. Or, certainly, one could get a strong combination of the two.
So what happens going forward? Is this a broken, or meaningless indicator? Perhaps, using a revised version of the Fed Model could help provide some additional perspective. Instead of using 10-year Treasuries, we’ve also recreated the Fed Model using long-term investment grade bond yields. Here is the Fed Model with that adjustment:
This version doesn’t show the same extreme in undervaluation as our “traditional” model, but it’s still near the top of the range observed over the past few decades.
All things considered, these Fed models add to other pieces of evidence that suggest it may be a mistake to write off the bull market in US stocks just yet. While US markets are certainly overvalued using the long-term CAPE PEs as a guide, long-term momentum remains comfortably intact, and forward-looking economic models continue to show very low probability for a US recession in coming months. As we’ve discussed in the past, the most devastating bear markets over the past century have occurred when markets face a devastating combination of extreme overvaluation, deteriorating intermediate to long-term momentum, and a strong deterioration in forward-looking economic indices, such as the Conference Board leading indicators, showing high probabilities of future recession and earnings dislocation. Right now, two out of three of our indicators are in green territory. The Fed Models aren’t necessarily the so-called “end all and be all.” But, they’re matching with other indicators, which makes the models hard to ignore, especially since they’ve performed reasonably well as action triggers across several distinct economic periods during the past 50 years.